Since ChatGPT Ads launched in the Netherlands, the question we hear most from clients is to be expected: does it work?
The honest answer is that it’s too early to say, for us and for anyone else in Europe. Ads started showing here on 24 August, and self-serve access to Ads Manager opened on 31 August. A few weeks isn’t enough to know how a channel performs, especially for B2B, where a lead can take months to become a customer.
We’ve tested the platform on our own account and are now setting up the first campaigns for clients. This post covers what we’ve learned so far, what matters for SaaS, lead generation and e-commerce, and how we’re approaching the tests.
The short version: ChatGPT Ads is worth testing, but only with a realistic budget (the platform minimum is €15 a day), tracking you can trust and a clear idea of what success looks like.
How ChatGPT Ads works in Europe right now
ChatGPT Ads matches ads to the context and intent of a conversation. Advertisers describe the topics, problems and situations where their offer might be relevant. These context hints aren’t exact-match keywords. The system looks at what someone is trying to achieve, not at the literal words they type. OpenAI explains how ad selection works here.
In Europe, ads aren’t personalised at this moment. The signals are the conversation itself, general location and device type. OpenAI is adding features in stages, but not everything available in the US works in European accounts yet. Custom audiences, for example, aren’t available for European campaigns. Conversion-optimised bidding is, which makes your conversion data more important: the platform optimises towards whatever you tell it counts as a result. OpenAI describes the European rollout here.
This is what makes the platform interesting. Someone using ChatGPT rarely just searches for “the best software”. They explain their problem, compare options and add conditions like team size, the systems they already use, budget or a process that’s slowing them down. That gives a relevant offer a chance to appear while someone is still exploring, before they’ve settled on a shortlist.
Who sees the ads, and why that matters
Ads only show to people on ChatGPT’s Free and Go plans. Plus, Pro, Business, Enterprise and Education users don’t see ads.
For B2B, this is the first thing to understand. We expect a good share of the decision-makers our SaaS and lead generation clients want to reach use ChatGPT through a paid or company account, and those people won’t see an ad. Nobody knows yet how big that group is for a given market. It’s one of the first things a test needs to show, and it’s a good reason not to shift budget based on the platform’s potential alone.
For e-commerce the situation is different. Consumer shoppers are much more likely to use the free version, so the reachable audience is broader.
What it means for SaaS, lead generation and e-commerce
SaaS. The strength is the comparison moment: someone describing their stack and asking which tool fits. The challenge is the sales cycle. A lead that enters the funnel today may not become a customer for months, so we judge tests on qualified demos booked and revenue. We do not just judge clicks.
Lead generation. Cost per lead matters, but a cheap lead means very little if the enquiry doesn’t fit. This is where you need to feed back your actual outcomes, not just form fills. Otherwise you can’t tell whether ChatGPT Ads brings in the right people or just .. more people.
E-commerce. Product feed campaigns are available in Europe. They show products as sponsored cards under ChatGPT’s answer, with title, price and image pulled from your feed. The matching works differently from Google Shopping, because people describe a use case instead of typing a product name. A feed with only brand names and internal SKUs gives the system very little to match on, so product titles and attributes that describe what the product is for become more important.
What we learned from testing it ourselves
Because ChatGPT Ads was a completely new platform, the only way to get hands-on experience before launch was to test it on our own account. We ran a short campaign of about a week on a small budget.
The main lesson was about budget. The minimum daily budget for euro accounts is €15, and a budget close to that minimum doesn’t produce enough data in one week to judge anything. You learn how the setup works, but you can’t tell whether the channel works for you. If you want to test ChatGPT Ads properly, plan for real test costs from the start, not the smallest amount the platform allows.
The setup also feels different from Google Ads. You describe situations and problems instead of choosing keywords, so the ad and the landing page need to connect to a specific need. A generic offer next to a detailed conversation stands out even more than it does in search.
How we set up a test with clients
We’re now setting up the first client campaigns. We start with clients we already know well because we know their funnels, their tracking and what a good lead or order looks like for them. That makes it much easier to judge what ChatGPT Ads adds on top of what already works.
Before a campaign goes live, we agree three things with the client. First, a fixed test budget of at least the €15 a day platform minimum, and in practice more, so the test can teach us something. It stays separate from channels that already perform. Second, a test period long enough to see what happens after the click, not just the click itself. Third, stop criteria: what would make us pause, adjust or scale.
We’re also not publishing benchmarks. No official ones exist yet, the Dutch market is young, and results will differ a lot per business. We’d rather share a useful conclusion once we have real data than an exciting number that doesn’t mean much.
Before you test a new channel
The biggest mistake would be to judge ChatGPT Ads separately from the rest of your marketing. A new ad platform won’t fix unclear positioning, a weak offer, a poor landing page or slow sales follow-up. It adds another touchpoint to a system that needs to work as a whole.
ChatGPT Ads may help you show up earlier in the research process. Google Ads captures a different kind of active demand. LinkedIn, Meta and organic visibility, whether that sits with your team or a specialist, build familiarity and trust. Landing pages, email and sales follow-up decide how much of that attention turns into pipeline and profit.
So we work foundation before spend. We find it crucial to connect measurement to CRM and pipeline data. CPC and CTR are part of the picture, but lead quality, orders and commercial outcomes matter more. When the data shows ChatGPT Ads adds relevant demand, there’s a case for scaling. When the signals aren’t strong enough yet, we keep learning and adjust the test.
What we mean by a solid foundation
Before you add a new channel, your setup needs to be able to tell you whether it worked. We call that the foundation, and it comes down to three things:
- Tracking you can trust. Every conversion that matters (a demo request, a form, an order) is measured correctly, and only once, on every channel. If the numbers in your ad platforms don’t match what ends up in your CRM or shop, you can’t compare a new channel with the ones you already run.
- An agreed definition of a good result. Marketing and sales agree on what counts as a qualified lead or demo. For e-commerce, that means a profitable order, not just any order. Without that, “it worked” means something different to everyone.
- Actual outcomes flowing back into the platforms. What happens after the conversion gets sent back to the ad platforms: a demo that became a deal, an order that wasn’t returned. Conversion-optimised bidding learns from the data you give it. If you only send form fills, it looks for more form fills. If you send actual outcomes, it looks for more of the right people.
If one of these is missing, a ChatGPT Ads test will give you clicks and costs, but you won’t be able to tell whether it brought in business.
In a strategy call we check exactly this across your Google Ads, Analytics and Merchant Center setup. We run Google, Microsoft, LinkedIn and Meta campaigns every day, so if the foundation is in place, we’re happy to set up a ChatGPT Ads test with you too.
FAQ
Do my buyers even see ChatGPT Ads?
Only if they use ChatGPT on the Free or Go plan. Paid plans like Plus, Business and Enterprise are ad-free. For B2B that may exclude part of your audience, and for e-commerce the reach is broader. The only way to find out for your market is a controlled test.
What budget do I need to test ChatGPT Ads?
The platform minimum is €15 a day for euro accounts. The test budget might need to be in principle higher to get meaningful traffic during the test period, and kept separate from channels that already perform. We agree the budget, the test period and the stop criteria before launch.
Does ChatGPT Ads replace Google Ads?
No. It reaches people at a different moment, often earlier in their research. Google Ads still captures active search demand. We see ChatGPT Ads as an addition to be tested, not a replacement.
How do you measure ChatGPT Ads when the sales cycle is long?
By connecting it to CRM and pipeline data instead of stopping at clicks. For SaaS that means qualified demos booked and pipeline value, for lead gen the quality of enquiries, and for e-commerce orders and margin. That takes longer than a few weeks, which is why we aren’t publishing benchmarks yet.
